• Hardware Platform (Jetson / GPU) : Triton Inference server • DeepStream Version : 5.1
We are using Video analytics on multiple cameras on deepstream. for output we have a requirement of individual camera output for which we are using demultiplexer. however to view one camera output we need one node port on kubernetes GPU.
We have a limitation on the number of nodeports available. Can you tell me a alternative solution for this challenge
I want a way out in which possibly i could resuse the same nodeports for different cameras based on user request for live streaming. We have a limitations on the number of Nodeport availability in kubernetes, so we cannot increase the number of nodeports